Ncut在颅脑MRI肿瘤提取中的应用研究

ON APPLYING NORMALISED CUT TO BRAIN MRI TUMOUR EXTRACTION

颅脑肿瘤自身边缘包含重要的病变信息,提取脑肿瘤区域,对脑部疾病的诊断和治疗具有重要意义。Ncut(Normalized Cut)是基于图理的典型分割方法。将Ncut算法应用到颅脑MRI肿瘤图像的分割中,针对不同颅脑MRI肿瘤图像,进行相关参数测试,选择合适的权重及参数,进行颅脑MR肿瘤的提取。通过利用Matlab进行仿真测试可知Ncut方法能够提取出肿瘤所在的基本轮廓,取得较好效果。

The brain tumour edge contains important pathology information itself; it has the important meaning to extract brain tumour area for brain disease diagnosis and treatment. Neut algorithm is the typical segmentation method based on graph theory. We apply the Ncut algorithm to MRI brain tumour image segmentation, test the related parameters according to different brain MRI tumour images, and select appropriate weights and parameters to extract MR image brain tumour. Through the use of Matlab in simulation test, we know that the Ncut method can extract the basic outline of tumour and achieve good effect.